Saffron thistle, Carthamus lanatus subsp. lanatus and baeticus (Asteraceae), is the most abundant thistle in southern Australia, where it primarily invades cereal crops and grazing areas. A classical biological control program for saffron thistle began in the early 1990s, with a series of surveys in the Mediterranean regions of Europe and regular monitoring of study sites in France. Two fly species, Urophora mauritanica that attacks flower heads and Botanophila turcica that feeds on rosette crowns, were selected as the most promising insects for biological control and studied in more detail. The rust fungus Puccinia sommieriana and the leaf-spot fungus Septoria centrophylli were the only pathogens considered for preliminary host-specificity testing. The two fungi and the fly U. mauritanica were found to be capable of attacking safflower cultivars, a closely related crop species to saffron thistle, and were thus not pursued further. The rosette crown-feeding fly B. turcica was the only candidate agent that did not attack safflower. Field studies showed that this fly could kill saffron thistle rosettes but in most instances attacked plants were able to compensate and continue to grow, although considerably less seed was produced. Nonetheless, questions were raised about the ability of this candidate agent to reduce populations of saffron thistle. Consequently, comprehensive host-specificity testing was not performed and it was not released in Australia. Research has also been exploring the potential of the inundative approach (bioherbicide) for saffron thistle control using a Phomosis sp. fungus that already occurs in Australia, and is ongoing.

Chondrilla juncea, skeleton weed, is a herbaceous perennial of Eurasian origin, first recorded in Australia in 1917. It spread throughout the south-eastern Australian wheat belt from the 1920s to 1940s, where it dramatically reduced yields, increased production costs and proved difficult to control. By the 1950s it was considered Australia’s worst weed. It has a deep perennial rootstock which produces rosettes from autumn to spring and a multi-branched, wiry, aerial stem during summer. This bears many capitula and the seeds, each bearing a pappus, are readily dispersed by wind. Once established in an area, density increases by regeneration from lateral roots and from segments of the rootstock when damaged, e.g. by cultivation. Densities in excess of 100 rosettes per m² were common. C. juncea is an apomict and exists as a number of genetically discrete forms, three of which occur in Australia. One of these was by far the most widespread and economically important. Detailed studies in Mediterranean Europe eastwards to Iran examined the impact of natural enemies as part of the weed’s overall ecology and examined nine species in some detail. Two were discarded as insufficiently host-specific, three more because of unresolved issues over specificity and four were introduced, three of which established widely. Of these, the rust Puccinia chondrillina, the first plant pathogen deliberately introduced for weed control, was extremely successful, dramatically reducing the density of the weed over a few years. The eriophyid mite Eriophyes chondrillae and the gall midge Cystiphora schmidti also established widely and assisted in weakening the plant. The economic benefit of this program has been calculated as $1.426 billion in 2004–05 dollars with a benefit cost:ratio of 112:1. The strains of the rust and the mite initially introduced attacked only the common widespread form of the weed and the other two forms gradually spread over most of the weed’s distribution. Later introductions of additional strains of P. chondrillina and E. chondrillae for the other two forms did not provide effective control, but the importance of those forms never reached that of the common form and skeleton weed is no longer considered a significant problem.

Chrysanthemoides monilifera invades native vegetation of southern Australia and is a major threat to the integrity of natural ecosystems. Two subspecies have mostly allopatric distributions: boneseed (subsp. monilifera) occurs throughout south-eastern Australia and bitou bush (subsp. rotundata) is mainly restricted to coastal habitats in NSW. Twelve potential agents have been studied, and some released, as part of a biological control program for C. monilifera. Four agents are established on bitou bush (the bitou tip moth Comostolopsis germana, the leaf-roller moth Tortrix sp., the bitou seed fly Mesoclanis polana and the bitou tortoise beetle Cassida sp. 3) and affect certain plant growth parameters, but have not yet caused substantial decline of infestations. One other agent (the boneseed leaf buckle mite Aceria sp.) is possibly established on boneseed. Four agents have failed to establish (three leaf beetles Chrysolina scotti, C. sp. B and C. picturata and the lacy-winged seed fly Mesoclanis magnipalpis). The bitou leaf beetle Ageniosa electoralis was rejected. Testing of the systemic rust fungus Endophyllum osteospermi is in progress and testing of the boneseed seed fly Mesoclanis dubia is incomplete.

Cryptostegia grandiflora, rubber vine, was introduced into Australia before 1875 and eventually became a major weed in northern Qld and a Weed of National Significance. It grows as a shrub or a scrambling vine capable of smothering tall trees and can form monocultures in riparian situations. Efforts to find biological control agents in the native range consisted of surveys and preliminary host-specificity testing performed from 1985–88 at a field station in Toliare, Madagascar. Although only a limited arthropod fauna was found, several species were selected for study. Two agents, the leaf-feeding moth Euclasta whalleyi and the autoecious rust fungus Maravalia cryptostegiae, were subsequently released in Australia. E. whalleyi established throughout the range of rubber vine, though it was not detected until nine years after release. On occasion it can be abundant on rubber vine, causing significant damage. M. cryptostegiae established very rapidly throughout most of the weed’s range and is a very effective agent providing a high degree of control. An evaluation of the project estimated a benefit:cost ratio of $109 at a discount rate of 8%. Based on benefit:cost ratio, the rubber vine project was ranked third (behind those for prickly pear and skeleton weed) of all Australian weed biological control projects. The control of rubber vine is therefore one of Australia’s best successes.

Cytisus scoparius is a leguminous shrub that is a serious weed of cool climate areas in south-eastern Australia. Biological control for Australia was initiated because of large infestations in the Barrington Tops region of NSW. A biological control program against broom has been underway since 1993, carried out by several state and federal agencies, much of it in collaboration with other countries, particularly New Zealand. This program has resulted in the release of three insects and a mite in Australia. All the agents have established somewhere in Australia and a number are causing slight damage in southern NSW, Vic and Tas. No significant damage to broom has occurred at the Barrington Tops, but the mite was only released in this area in 2010. Biological control of broom has not been considered a success to date despite considerable ecological research suggesting natural enemies are important in the invasion dynamics of broom.

Water hyacinth, Eichhornia crassipes, a native of Amazonia, was first recorded in Australia in 1894. It is widespread in waterways of the east coast, particularly north of Sydney. Biological control of water hyacinth in Australia began in the 1970s and four agents have been released. Neochetina bruchi, N. eichhorniae and Niphograpta albiguttalis are widely established, whereas Xubida infusellus is known to be established at one site only. Successful control has been achieved at many locations. However, a range of factors such as flooding, drying, cold winters, herbicidal application etc. mitigate against successful biological control at other sites. There is a need to develop integrated management strategies to best utilise the established agents in combination with other control techniques, and there is a need for additional agents, some of which are being studied or have been released overseas. Of particular interest is a plant hopper, Megamelus scutellaris, recently released in the USA. Investing in biological control of water hyacinth in Australia and elsewhere, and in integrated management in South Africa, has been extremely cost-effective with benefit:cost estimates up to 27.5:1 in Australia and 124:1 in Benin, and savings of over $US200 per ha compared to herbicidal control in the USA.

Some of the seven Fumaria species introduced to Australia, especially F. densiflora, have become important weeds of cropping in south-eastern Australia. A survey initiated in part of the region of origin of most of the species, France, found two insect species and two plant pathogen species that could be of interest for use in bio-control. These species were absent in a survey of Fumaria species in Australia. Further research is needed on the weed status in Australia, plus additional surveys in the region of origin, and research on the ecology and host-range of potential agents for Fumaria species.

Opuntia spp. were major weeds in Qld and NSW. The biological control program that ran from 1911 to 1939 resulted in the release of 20 biological control agents, with 14 reported as established, some of which appear to have subsequently died out. Cactoblastis cactorum and Dactylopius spp. are considered to be the main agents that controlled Opuntia spp. following their release, and still do so today. This biological control program is considered to be the most successful biological control program in Australia. No Opuntia spp. in Qld or NSW is now considered to be a major problem. In 2005 dollar terms, a conservative estimate of the return on investment was $147.4 for each $1 spent, with a net present value of $3110.3 million.

Parthenium hysterophorus, an annual herbaceous plant native to the tropical Americas, is now a major crop and pasture weed in Australia as well as many parts of Africa and Asia. It is also a human health hazard, causing severe allergic dermatitis and respiratory problems. A major biological control program against parthenium in Qld started in 1977 with scientists based in Mexico and Brazil, and continued until 2002. Over 144 species of herbivorous insects and several pathogenic fungi were identified, of which nine insect species and two rust fungi were introduced into Qld. Seven species of insects and two rust fungi have established on parthenium and are widespread. Key agents are the stem-galling moth Epiblema strenuana, the leaf-feeding beetle Zygogramma bicolorata, the stem- and root-feeding weevil Listronotus setosipennis, the seed-feeding weevil Smicronyx lutulentus and the summer rust Puccinia xanthii var. parthenii-hysterophorae. The combined impact of these agents has resulted in significant reductions in the abundance and impact of parthenium in most situations and seasons, though serious infestations can still occur. Some of these agents have been successfully introduced into India, Sri Lanka and China, and are being considered for use in eastern and southern Africa.

Pistia stratiotes, water lettuce, is a free-floating aquatic weed that has naturalised in Qld and NSW. It is also found in the NT, where it is considered to be native. Water lettuce is fast-growing and has the ability to form dense mats, blocking waterways and interfering with access to water, fishing and hydro-electric flows. Biocontrol began in Australia in 1982, with the opportunistic release of the weevil Neohydronomus affinis from Brazil. This insect quickly established in most parts of Qld where it was released. Complete control in northern Qld was achieved in 18 months, while control in southern Qld was achieved in over two years. However, water lettuce infestations in south-western Qld are also actively treated mechanically or with herbicide, or are flushed out during floods, and this inhibits the maintenance of weevil populations. In NSW, the weevil was not released, as the weed is a target for eradication and is being actively controlled.

Raphanus raphanistrum (Brassicaceae), wild radish, is an annual winter-growing herb that is widespread in southern Australia where it is one of the economically most important weeds of annual cropping. Wild radish was surveyed for potential agents, insects and pathogens in part of the region of origin, southern Europe and North Africa. Potential agents, pathogens and the gall midge Gephyraulus raphanistri, were tested to assess their host-range, but none were sufficiently host-specific. Complementary surveys of pathogens, including soil bacteria, have been made in Australia to determine which organisms are already present and could be used in development of bioherbicides. Ants are an important cause of seed predation in both the regions of origin and introduction. Wild radish is very closely related to edible radish Raphanus sativus and the important crop, canola. This limits the possible range of biocontrol agents due to lack of host-specificity to the weed. Priorities for future research include examining low-dose herbicide–pathogen interactions, examining the species boundaries and host-specificity of the downy mildew Hyaloperonospora parasitica, developing pathogens already in Australia as mycoherbicides, and promoting beneficial pathogens and rhizobacteria through stubble and soil management.

European blackberry, a group of closely related species belonging to the Rubus fruticosus aggregate (Rosaceae), is an introduced invasive shrub of agriculture, forestry and natural ecosystems in Australia. This chapter describes the various Australian biological control projects targeting European blackberry carried out over the years and summarises similar efforts undertaken in other countries. Surveys in Europe in the 1970s identified three potential agents for European blackberry, the stem-boring sawfly Hartigia albomaculata, purple blotch fungus Septocyta ruborum and leaf-rust fungus Phragmidium violaceum. The purple blotch fungus was not investigated at the time and preliminary host-specificity tests with the sawfly showed that larvae were able to feed on a number of cultivated blackberry and rose varieties. In 1984, while efforts were underway in Europe to select isolates of the leaf-rust fungus, it was found in Australia after an unauthorised introduction. Nonetheless, one of the selected and tested isolates (F15) was officially released in Australia in the early 1990s. Subsequent research addressed knowledge gaps considered critical to the biological control success of European blackberry, including a taxonomic revision of the R. fruticosus aggregate in Australia, climate modelling to predict impact of the rust fungus, and studies of pathogenicity and population genetics of the fungus to explain variation in disease intensity across the landscape. In 2000, eight additional and genetically diverse isolates of the rust fungus were sourced using a trap garden established in Europe that comprised blackberry clones from Australia. Following demonstration of host-specificity and approval for release in 2004, a large-scale release program, including re-release of F15, was established and the fate of the additional isolates monitored using molecular markers. Concurrently, surveys for other potential candidate agents were performed in Europe and as a result the purple blotch fungus is being investigated in more detail.

Salvinia molesta, salvinia, is a sterile floating fern that has spread to most tropical and subtropical countries. Rapid growth rates allow it to blanket still or slow-moving water bodies very quickly, causing environmental, sociological and economic problems. A small black weevil, Cyrtobagous salviniae, from Brazil, which is specialised on Salvinia species, was released in Australia in 1980 on Lake Moondarra, Mt Isa and provided extraordinary control of the weed within 15 months. Thereafter, releases elsewhere in Australia and in Papua New Guinea, Sri Lanka and many other countries resulted in extremely high levels of control, mostly within three years. There are situations where the weevil is not effective, such as on multi-layered salvinia mats or when it grows as an understorey plant, but integrated management strategies that include biological control can overcome these restrictions. Recent studies have confirmed the value of this agent in temperate regions. Other control methods for salvinia are impractical, ineffective and costly, whereas biological control has provided long-term sustainable management with very positive benefit:cost ratio estimates from 11 to 53:1. The control of salvinia by the salvinia weevil is a classic contemporary example of the benefits of biological control of weeds.

Sicklepod Senna obtusifolia and its congener foetid senna Senna tora are serious weeds of field and pasture in many tropical and subtropical regions such as Australia, the USA and the Pacific islands. In Australia, sicklepod infests over 60 000 ha of the tropical eastern coast of Qld from Cape York to Mackay. However eco-climatic modelling suggests a larger potential range.

Classical biological control was considered by the Commonwealth Institute of Biological Control for several island nations while American scientists researched mycoherbicides for use in cropping situations along the Gulf Coast of the USA. Brazilian scientists also investigated possibilities for Brazil. Surveys were then conducted in Mexico and Central America for Qld from 1991 to 1993 and 108 species representing six insect orders and 40 families were found.

Two insects were studied in quarantine in Australia. The psyllid Mitrapsylla albalineata was distributed widely in Mexico and infested terminal leaves. However, it utilised several native Australian plant species, bred successfully through three generations on nine Senna spp., and was considered too problematic for release in Australia. The second, a weevil Conotrachelus sp., entered a reproductive diapause after being reared for five generations. It ceased oviposition and the colony was lost before host testing could be undertaken. Mixed populations of four bruchid seed-feeding beetles Sennius fallax, Sennius lebasi, Sennius morosus and Sennius rufomaculatus were commonly found in Mexico. They are not monophagous, but have host-ranges mostly confined to Senna spp. Although it is unlikely that they would be approved for release in Australia under present guidelines, they may warrant further study. The hesperiid Typhedanus undulatus, the pierid Eurema salome and the pyralid Anabasis ochrodesma, all from Mexico, may warrant further study.

The project was terminated prematurely because of funding issues. No agent has been released and sicklepod remains a serious concern in northern Qld.

Silybum marianum is one of the most widespread exotic weedy carduine thistles in Australia. In Vic it infests more than 4.7 million ha and vast tracts of NSW are also affected. The only natural enemy released as a biological control agent against it in Australia is the thistle receptacle weevil Rhinocyllus conicus. This weevil is probably established in Vic but this requires confirmation. Two rust species may warrant further investigation as potential agents. If more agents are required, exploration should be focused in south-western Spain and northwestern Africa.

Sowthistles, Sonchus spp. (Asteraceae), are widely known as serious weeds affecting crop production in many regions of the world. Australia has three species, the weedy introduced Sonchus oleraceus and Sonchus asper and the native Sonchus hydrophilus. S. oleraceus is considered a major weed of the south-east Qld and northern NSW cropping region, whereas S. asper has relatively minor importance in Australia. Surveys of S. oleraceus identified several potential biocontrol agents already present in Australia including Miyagia pseudosphaeria, Aceria thalgi and the potential mycoherbicide pathogen Aschochyta sonchi. In Australia, A. thalgi and M. pseudosphaeria are widespread and could be used in a conservative approach to biological control of exotic Sonchus spp. Potential agents identified in preliminary surveys in the native distribution in France could be considered for a classical biocontrol approach. Host-range studies would need to take into account the close relationship between introduced and native Sonchus in Australia.

Tribulus terrestris, caltrop, is an annual summer-growing prostrate herb that is widespread in Australia. Its burrs contaminate agricultural produce and the plant is toxic to livestock. It is an approved target for biocontrol, but no introductions of biocontrol agents have been made into Australia. Two organisms, the seed-feeding moth Ephysteris subdiminutella ferritincta and an unidentified eriophyid mite similar to Aceria tribuli causing galls on leaves and stems, are already present in Australia and damaging on the plant. The main biocontrol agents used overseas, Microlarinus weevils, have a history of spreading between islands and could potentially arrive in Australia from New Guinea where at least one species is established. Issues that will need to be resolved before the application of biocontrol include determining the native or introduced status of T. terrestris, the host-specificity of potential agents in the context of the risk to the diverse flora of native Tribulus species, and the lack of economic assessment showing the costs associated with the weed.

The annual herb, Xanthium occidentale, Noogoora burr, was one of the first weeds targeted by biocontrol worldwide. In the 1950s it was considered one of the worst weeds in Australia, owing largely to its impact on the wool industry. In Australia it was the focus of major efforts both prior to and following World War II; these resulted in the release of three biocontrol agents, two stem-boring beetles and a seed-feeding fly. In 1974 the rust fungus Puccinia xanthii was introduced through unknown means, and in 1982 the stem-galling moth Epiblema strenuana was released against Parthenium hysterophorus but it also attacked Noogoora burr. The rust caused almost complete control across much of Australia, except in more arid regions and in the wet-dry tropics. The moth may also have provided some control, but the remaining three insects had no impact. Recent work has focused on finding pathogenic rust isolates that are better adapted climatically to conditions in northern Australia, but it has so far not been successful. Nonetheless, the benefits of the biocontrol program (including the rust) were estimated to exceed $A23 million in 2004–05, and Noogoora burr is now considered a relatively minor problem nationally.

Arum lily Zantedeschia aethiopica is an invader of native wetlands and wet pastures of the lower south-west of southern Australia that has proved difficult to control by conventional means. The flowers are harvested from weed populations and the plant is used as an ornamental although it is declared noxious in WA. The weed originates from southern Africa, where initial surveys for bio-control agents focused on seed-feeding insects because of the potential conflict of interest with the plant’s use in horticulture. Suitable organisms for use in biocontrol were not found. Surveys were then made in WA for disease-causing organisms that could be used in an inundative approach to biocontrol. Potential plant pathogens were found or identified from the literature, but have not been studied further. Future work should focus on completing surveys for pathogens in Australia and investigating the potential for mycoherbicide development, especially for organisms associated with the rhizome, the main perennating organ for arum lily.